woshin

/

FLUX.1-Canny-dev-redux

like 8

I have two quick questions about FLUX.1-Canny-dev-redux:

1. Was the output from FLUX.1-Redux-dev used as an additional condition or input during fine-tuning?

2. How much training data was used for this fine-tuning (e.g., number of images or prompts)?

Thanks in advance!